search

HUANGLIZI / LViT

[IEEE Transactions on Medical Imaging/TMI] This repo is the official implementation of "LViT: Language meets Vision Transformer in Medical Image Segmentation"
MIT License
283 stars 26 forks source link

Question about UNet #46

Closed blz822 closed 3 months ago

blz822 commented 3 months ago

Dear Zihan Li,

Thank you for your excellent work!

I have been following your work and have successfully used the setting 'model_type="LViT"' to run the code. However, I have encountered some challenges in replacing LVIT with U-Net for model training.

As discussed in Section 2.1 of your documentation: image

Your guidance on these matters would be greatly appreciated, as it would greatly assist me in understanding and utilizing your work more effectively.

Thank you for your time and consideration. I am looking forward to your valuable insights.

Best regards, Bao

HUANGLIZI commented 3 months ago

You can use the UNet we provided for the pre-training: https://github.com/HUANGLIZI/LViT/blob/main/nets/UNet.py. And you just need to replace the LViT with the UNet.